Pith. sign in

REVIEW 2 major objections 7 minor 27 references

Parameter Error Analysis for the 3D Modified Leray-alpha Model: Analytical and Numerical Approaches

T0 review · 2 major / 7 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read For a turbulence model with an unknown length scale, continuous data assimilation recovers the true flow exponentially up to an error proportional to the squared parameter mismatch.

desk verdict The ML-alpha parameter error theorem is promising but the proof has a genuine algebraic error in the central Young inequality; repairable but not ready. read the letter →

arxiv 2411.16324 v1 pith:PPOQKUOF submitted 2024-11-25 math.NA cs.NAmath-phmath.APmath.MP

classification math.NAcs.NAmath-phmath.APmath.MP MSC 35Q3576D0576F6565M70
keywords continuousdataassimilationmodifiedLeray-alphamodelparametererroranalysisturbulencemodelsnudgingwell-posednessestimates3DNavier-Stokesregularization
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper asks whether a turbulence model can still track the true flow when one of its physical parameters is unknown. For the three-dimensional viscous modified Leray-$\alpha$ model, the authors replace the true length scale $\alpha$ by a guessed value $\beta$ inside a continuous data assimilation system that receives sparse measurements of the true velocity. They prove that the assimilated solution converges exponentially to the true solution up to a time-independent error floor, provided the nudging strength, the measurement spacing, and the true solution's energy satisfy three explicit inequalities. The error floor grows with the squared difference of the squared parameters, $|\beta^2-\alpha^2|^2$, so it vanishes when the guess is correct. A sympathetic reader would care because this turns parameter uncertainty into a controlled error rather than a fatal flaw.

What carries the argument

The load-bearing object is the error equation for $g=w-u$, obtained by subtracting the true ML-$\alpha$ system (18) from the assimilated system (25). The difference in the filtered velocities reads $z-v = g+\beta^2Ag+(\beta^2-\alpha^2)Au$, so the unknown parameter enters as a forcing term proportional to $(\beta^2-\alpha^2)Au$ and its time derivative. The proof bounds each term of the resulting differential inequality with the Gagliardo-Nirenberg inequality and the interpolation property (4) of the coarse-mesh operator $I_h$, then applies a generalized Gronwall lemma (Lemma 1) that admits a time-averaged forcing. Conditions (1)--(3) are exactly the inequalities that keep the dissipative coefficients positive, turning the error dynamics into exponential decay plus a constant inherited from the time-averaged parameter-mismatch term.

What would settle it

Run the same spectral assimilation with a fixed $\alpha$ and two different $\beta$ values and measure the long-time error plateau; Theorem 3 predicts the plateau scales as $|\beta^2-\alpha^2|^2$, so a plateau that fails to shrink quadratically as $\beta$ approaches $\alpha$ would falsify the error mechanism. A second check: for a more energetic true solution (larger $M_1$) or smaller $\alpha$, search for any $(\eta,h,\beta)$ satisfying all three conditions; if none exists, the recovery guarantee is vacuous in that regime.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is Theorem 3: if conditions (1)--(3) hold, the error $g(t)=w(t)-u(t)$ between the assimilated unfiltered velocity and the true one obeys $|g(t)|^2+\beta^2\|g(t)\|^2 \le e^{-\lambda\nu t/2}(|g(0)|^2+\beta^2\|g(0)\|^2) + M_\alpha (e/(e^{1/2}-1))$, with $M_\alpha$ of order $|\beta^2-\alpha^2|^2$. The first term decays exponentially at a rate set by the viscosity and the box size; the second term is the price of not knowing $\alpha$, and it depends on the parameter mismatch through $M_\alpha$ defined in (46). The paper also proves the assimilated system (25) is globally well-posed and depends continuously on its initial data (Theorem 2). Interpreted plainly: with sufficiently strong nudging and sufficiently fine measurements, the guessed-parameter model locks onto the true solution, and whatever residual error remains is controlled by how wrong the guess was.

Load-bearing premise

The load-bearing premise is that some choice of nudging strength $\eta$, measurement spacing $h$, and guess $\beta$ satisfies all three inequalities in Theorem 3; the paper does not prove such a choice exists for a given flow, and only checks it in one numerical example.

Editorial extensions

If this is right

  • When $\beta=\alpha$, the parameter-mismatch constant $M_\alpha$ vanishes and the error decays to zero exponentially: the guessed model synchronizes exactly with the true solution.
  • For a misspecified $\beta$, the asymptotic error is bounded by a time-independent constant, so the assimilated solution remains a reliable approximation indefinitely rather than drifting away.
  • The three conditions translate into a tuning recipe: the nudging $\eta$ must exceed a threshold set by the true solution's energy and $\alpha$, while staying below bounds that tighten as the measurement spacing $h$ shrinks.
  • The continuous-dependence result makes the assimilated system well-posed and stable with respect to its initial condition, ensuring the recovery statement is not an artifact of a special start.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A practical parameter-recovery scheme is implicit: run the assimilation with several candidate $\beta$ values and choose the one with the smallest asymptotic error floor, since the floor is quadratic in the parameter gap; the paper does not implement this.
  • Because condition (1) demands a large $\eta$ while conditions (2)--(3) bound $\eta$ from above by terms involving $1/h^2$ and $1/h^4$, the admissible region is likely largest for moderate-Reynolds flows; a parameter sweep over $\alpha$ and $M_1$ would map where the guarantee applies.
  • The same filtered-velocity-difference mechanism suggests the error analysis carries over to other $\alpha$-models, where the parameter gap enters the filtered velocity analogously; the paper only lists this as future work.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 7 minor

Summary. The paper studies continuous data assimilation for the 3D modified Leray-alpha model when the lengthscale parameter alpha is unknown and replaced by a guess beta. It proves global well-posedness for the assimilated system and, under three sufficient conditions, an exponential-in-time recovery estimate for the error g = w - u, with asymptotic error M_alpha = O(|beta^2 - alpha^2|^2). Numerical simulations with the Dedalus package illustrate convergence when the conditions hold and divergence when the nudging parameter is too small.

Significance. If correct, this is a useful contribution to parameter-robust data assimilation for subgrid-scale turbulence models, extending the parameter-recovery framework of Carlson, Hudson, and Larios to a 3D alpha-model with explicit constants. The proof strategy is transparent and the numerical validation is a plus. However, the central proof contains a quantitative error in a Young inequality that invalidates the stated sufficient condition; the theorem is likely repairable, but the manuscript in its current form does not establish the main estimate.

major comments (2)
  1. [3.2, estimate (iii)] The displayed Young inequality in estimate (iii) is algebraically false. For the product c|g|^{1/2}\|g\|^{3/2}\|u\|, optimizing over the bound A\|g\|^2 + K\|u\|^4|g|^2 yields A^3K = 27/256; with K = 153c^4/(2^{11}\nu^3), the minimal admissible A is (3/4)(2^{11}/153)^{1/3}\nu \approx 1.122\nu, not 2\nu/5. Consequently the |g|^2-coefficient and condition (1) of Theorem 3 are not established by the proof as written, and the exponential recovery estimate does not close. The gap is repairable by replacing 153/2048 with 3375/2048 and strengthening the hypotheses accordingly, but this is a load-bearing correction.
  2. [Theorem 2, display (35)] The uniform bound \|z_m\|^2_{L^\infty([0,T];\dot V')} \le E_1(T) stated after (33) is not a consequence of (33). For a Fourier mode with eigenvalue \lambda, the ratio \|z_m\|_{\dot V'}^2/(|w_m|^2+\beta^2\|w_m\|^2) equals (1+\beta^2\lambda)^2/(\lambda(1+\beta^2\lambda)) = \lambda^{-1}+\beta^2, which can exceed 1 when \lambda is small and \beta<1. The subsequent Aubin-Lions compactness argument can be repaired with a constant depending on \lambda_1 and \beta, but the displayed inequality as written is incorrect.
minor comments (7)
  1. [Introduction] There are typographical errors: "Helmoltz" should be "Helmholtz" and "repectively" should be "respectively".
  2. [Lemma 2] The statement of Lemma 2 has a malformed norm ("sup |f(s)\|_L") and does not explicitly assume f \in L^\infty([0,\infty);H), although this is needed for M_1 to be finite and for the long-time estimates in Theorem 3.
  3. [Theorem 3] The theorem statement should explicitly include the assumptions f \in L^\infty([0,\infty);H) and u_0 \in V, since the proof of the M_\alpha bound uses sup_{s\ge 0}|f(s)| and Lemma 3, both of which require these hypotheses.
  4. [Lemma 1] Lemma 1 is cited to the companion paper [3] rather than proved; a short proof would make the paper self-contained and avoid relying on an unpublished reference.
  5. [Section 4.1] The reported value C_1 = 0.00739 appears inconsistent with the stated constants (c = \sqrt{3}, c_1 = \sqrt{32}, c_2 = 2, \nu = 0.75, \alpha = 0.3, M_1 = 0.00339); using c = \sqrt{3} in the formula gives approximately 0.0030. The numerical validation should be checked against the corrected theoretical constant.
  6. [Section 4.3] The text mentions "16 bit floating point value"; double precision is typically 64-bit, so this is likely a typo.
  7. [Theorem 3, hypotheses] The paper does not discuss whether conditions (1)-(3) are mutually satisfiable for a nontrivial range of parameters; since (1) is a lower bound and (2)-(3) are upper bounds on \eta, a remark on admissible parameter ranges would strengthen the applicability claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No material circularity: the recovery estimate is derived from first-principles energy estimates with explicit constants, and the self-citations are non-load-bearing.

full rationale

Theorem 3's error bound is not an input in disguise: the proof subtracts (18) from (25), decomposes B(z,w)-B(v,u), bounds the resulting terms via Gagliardo-Nirenberg, interpolation, and Young inequalities, and then applies a Gronwall lemma. The constants in the final estimate, including the decay rate gamma = nu*lambda_1/2 and the mismatch constant M_alpha in (46), are explicit expressions in the physical parameters and norms; none is fitted to data or renamed from an empirical quantity. Conditions (1)-(3) are stated sufficient hypotheses, not conclusions forced by the numerics, and the numerical section only selects parameters satisfying those hypotheses to illustrate convergence. The only self-citations are Lemma 1, whose proof is referred to the authors' companion paper [3], and the proof pattern of Lemma 2, which mirrors a lemma in [3]. Both are standard, independent estimates (an elementary Gronwall variant and an energy estimate using (B(v,u),u)=0); neither assumes the target error bound, so they are not load-bearing circularity. The reviewer-flagged issue in estimate (iii), concerning the constant in a Young inequality application, would be a correctness or proof-gap concern rather than a circular reduction, and thus does not change the circularity verdict.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

No free parameters are fitted. The central error bound is a parameter-free derivation in the sense that all constants are explicit or from cited inequalities. The paper relies on prior well-posedness of the ML-alpha model, the approximation property of the interpolant, and a Gronwall lemma from the authors' companion paper [3]; none of these assume the target result.

assumptions (3)
  • domain assumption Global well-posedness of the 3D modified Leray-alpha model (Theorem 1), cited from Ilyin, Lunasin, Titi 2006.
    Invoked at the start of Section 3 to ensure the existence of the regular solution u used in the error analysis.
  • domain assumption The linear interpolant I_h satisfies the approximation property (4) with constants c1, c2.
    Used throughout the proofs of Theorems 2 and 3 to control the nudging error.
  • standard math Alternative Gronwall inequality (Lemma 1) with the stated constants, cited from the authors' companion paper [3].
    Used at the end of Theorem 3 to convert the differential inequality into the exponential decay bound. The lemma is not proved in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Parameter Error Analysis for the 3D Modified Leray-alpha Model: Analytical and Numerical Approaches." pith.science (2026). https://pith.science/paper/PPOQKUOF

@misc{pith2026241116324,
  author       = {Pith},
  title        = {Pith review of: Parameter Error Analysis for the 3D Modified Leray-alpha Model: Analytical and Numerical Approaches},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PPOQKUOF}},
  note         = {Machine review of arXiv:2411.16324}
}
abstract

In this study, we conduct a parameter error analysis for the 3D modified Leray-$\alpha$ model using both analytical and numerical approaches. We first prove the global well-posedness and continuous dependence of initial data for the assimilated system. Furthermore, given sufficient conditions on the physical parameters and norms of the true solution, we demonstrate that the true solution can be recovered from the approximation solution, with an error determined by the discrepancy between the true and approximating parameters. Numerical simulations are provided to validate the convergence criteria.

Figures

Figures reproduced from arXiv: 2411.16324 by the authors.

Figure 1
Figure 1. Error plot of modified Leray-α model with high η value-without random initial conditions case. as well as check for the hypotheses given in Theorem 3, i.e. (1) 3η 4 − 153M2 1 c 4 2 11ν 3α4 > 0 ⇐⇒ η > C1 := 4 3 · 153M2 1 c 4 2 11ν 3α4 , (2) ηc2 1h 2 + 5η 2β 2 2ν c 2 1h 2 + 303 c 4M2 1 4 4ν 3α4 − ηβ2 < ν 4 , (3) ηc2 2h 4 + 5c 2 2 η 2β 2h 4 2ν < νβ2 4 . Here, the constants are c = 4 3 √ 3 3/4 [17], c1 = √ 32, and c2 = … view at source ↗
Figure 2
Figure 2. Velocity contour of modified Leray-α model with high η value-without random initial conditions case at t = 0 [PITH_FULL_IMAGE:figures/full_fig_p020_2.png] view at source ↗
Figure 3
Figure 3. Velocity contour of modified Leray-α model with high η value-without random initial conditions case at t = 400. system (18) has a random component which is u = (u0, v0, w0) : u0 = 0.05 sin 2πxz + 0.02 ∗ X − 0.01, v0 = 0.05 sin 2πxy + 0.02 ∗ X − 0.01, w0 = 0.05 sin 2πyz + 0.02 ∗ X − 0.01, where X is a random variable drawn from a uniform distribution [PITH_FULL_IMAGE:figures/full_fig_p020_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Error plot of modified Leray-α model with low η value-without random initial conditions case [PITH_FULL_IMAGE:figures/full_fig_p021_4.png]
Figure 5
Figure 5. Figure 5: Velocity contour of modified Leray-α model with low η value-without random initial conditions case at t = 0. The initial conditions for the assimilated model (25) is taken to be w = (ˆu0, vˆ0, wˆ0) : uˆ0 = 0.05 sin πx cos πy, vˆ0 = 0.05 sin πy cos πz, wˆ0 = 0.05 sin πz…
Figure 6
Figure 6. Figure 6: Velocity contour of modified Leray-α model with low η value-without random initial conditions case at t = 400 [PITH_FULL_IMAGE:figures/full_fig_p022_6.png]
Figure 7
Figure 7. Figure 7: Error plot of modified Leray-α model with high η value-with random initial conditions case. Here, we have ν = 0.75 and α = 0.3. M1 = 0.00355, h = 0.043, and β = 0.35. We compare the results when η = 1.5 > C1 ≈ 0.00811 (results in 7-9) and η = 0.0001 < C1 ≈ 0.02626 (res…
Figure 8
Figure 8. Figure 8: Velocity contour of modified Leray-α model with high η value-with random initial conditions case at t = 0 [PITH_FULL_IMAGE:figures/full_fig_p023_8.png]
Figure 9
Figure 9. Figure 9: Velocity contour of modified Leray-α model with high η value-with random initial conditions case at t = 400. 4.3. Discussions on the numerical computations. When explor￾ing the behavior of the ML-α data assimilation model, it is found that convergence is always achieve…
Figure 10
Figure 10. Figure 10: Error plot of modified Leray-α model with low η value-with random initial conditions case [PITH_FULL_IMAGE:figures/full_fig_p024_10.png]
Figure 11
Figure 11. Figure 11: Velocity contour of modified Leray-α model with low η value-with random initial conditions case at t = 0. truncation error that occurs in the 16 bit floating point value. When η is set to 1.5, the convergence of the original and assimilated system reach an order of ma…
Figure 12
Figure 12. Figure 12: Velocity contour of modified Leray-α model with high η value-with random initial conditions case at t = 400. these same conditions with an additional randomization term added to the original system. This serves to help us evaluate the role of random perturbations in t…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 27 canonical work pages

  1. [7]

    Parameter recovery for the 2 dimen- sional Navier-Stokes equations via continuous data assimilation, SIAM Journal on Scientific Computing, v.42, A250-A270, 2020

    Carlson, E.; Hudson, J.; Larios, A. Parameter recovery for the 2 dimen- sional Navier-Stokes equations via continuous data assimilation, SIAM Journal on Scientific Computing, v.42, A250-A270, 2020

  2. [3]

    Parameter Analysis in Continuous Data Assimilation for Various Turbulence Models

    Albanez, D.A.F.; Benvenutti, M.J.; Little, S.; Tian, J. Parameter Analysis in Continuous Data Assimilation for Various Turbulence Models, arXiv:2409.03042, 2024

  3. [1]

    Continuous data assimilation for the three-dimensional Navier-Stokes-α model, Asymp- totic Analysis, v.97, 139-164, 2016

    Albanez, D.A.F.; Nussenzveig Lopes, H.J.; Titi, E.S. Continuous data assimilation for the three-dimensional Navier-Stokes-α model, Asymp- totic Analysis, v.97, 139-164, 2016

  4. [2]

    Continuous data assimilation algo- rithm for simplified Bardina model, Evolution Equations and Control Theory, v.7, 33-52, 2018

    Albanez, D.A.F.; Benvenutti, M.J. Continuous data assimilation algo- rithm for simplified Bardina model, Evolution Equations and Control Theory, v.7, 33-52, 2018

  5. [4]

    Continuous data assimilation using gen- eral interpolant observables, Journal of Nonlinear Science, 24, 277-304, 2014

    Azouani, A.; Olson, E.; Titi, E.S. Continuous data assimilation using gen- eral interpolant observables, Journal of Nonlinear Science, 24, 277-304, 2014

  6. [5]

    Continuous data assimilation with stochastically noisy data, Nonlinearity, 28, p.729, 2015

    Bessaih, H.; Olson, E.; Titi, E.S. Continuous data assimilation with stochastically noisy data, Nonlinearity, 28, p.729, 2015

  7. [6]

    On the Clark-α model of turbulence: global regularity and long-time dynamics, Journal of Turbulence, 6, N20, 2005

    Cao, C.; Holm, D.D.; Titi, E.S. On the Clark-α model of turbulence: global regularity and long-time dynamics, Journal of Turbulence, 6, N20, 2005

  8. [8]

    On a Leray–α model of turbulence, Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 461, 629-649, 2005

    Cheskidov, A.; Holm, D.D.; Olson, E.; Titi, E.S. On a Leray–α model of turbulence, Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 461, 629-649, 2005

Show all 27 references
  1. [9]

    Ev ans, L.C.Partial Differential Equations, American Mathematical Society, v.19, 2022

  2. [10]

    Continuous data assimilation for the 2D Bénard convection through velocity measurements alone, Physica D: Non- linear Phenomena, v.303, 59-66, 2015

    F arhat, A.; Jolly, M.S.; Titi, E.S. Continuous data assimilation for the 2D Bénard convection through velocity measurements alone, Physica D: Non- linear Phenomena, v.303, 59-66, 2015

  3. [11]

    F arhat, A.; Lunasin, E.; Titi, E.S. Abridged continuous data assimilation for the 2D Navier–Stokes equations utilizing measurements of only one compo- nent of the velocity field, Journal of Mathematical Fluid Mechanics, v.18, 1-23, 2016

  4. [12]

    Data assimilation algorithm for 3D Bénard convection in porous media employing only temperature mea- surements, JournalofMathematicalAnalysisandApplications, v.438, 492-506, 2016

    F arhat, A.; Lunasin, E.; Edriss S; Titi, E.S. Data assimilation algorithm for 3D Bénard convection in porous media employing only temperature mea- surements, JournalofMathematicalAnalysisandApplications, v.438, 492-506, 2016

  5. [13]

    F arhat, A.; Lunasin, E.; Titi, E.S. On the Charney conjecture of data assimilation employing temperature measurements alone: the paradigm of 3D planetary geostrophic model, MathematicsofClimateandWeatherForecasting, v.2, 2016. PARAMETER ANALYSIS FOR 3D MODIFIED LERAY-ALPHA MODEL 27

  6. [14]

    Foias, C.; Holm, D.D.; Titi, E.S. The three dimensional viscous Camassa- Holm equations, and their relation to the Navier-Stokes equations and turbu- lence theory, Journal of Dynamics and Differential Equations, v.14, 1-35, 2002

  7. [15]

    Navier-Stokes Equations and Turbulence, Cambridge University Press, 2001

    Foias, C.; Manley, O.; Rosa, R.; Temam, R. Navier-Stokes Equations and Turbulence, Cambridge University Press, 2001

  8. [16]

    Partial Differential Equations, Dover Publications, Inc., New York, 2008

    Friedman, A. Partial Differential Equations, Dover Publications, Inc., New York, 2008

  9. [17]

    An Introduction to the Mathematical Theory of the Navier-Stokes Equations, Steady-State Problems, Springer New York, 2011

    Galdi, G.P. An Introduction to the Mathematical Theory of the Navier-Stokes Equations, Steady-State Problems, Springer New York, 2011

  10. [18]

    Fluctuation effects on 3D-Lagrangian mean and Eulerian mean fluid motion, Physica D: Nonlinear Phenomena, 215–269, 1999

    Holm, D.D. Fluctuation effects on 3D-Lagrangian mean and Eulerian mean fluid motion, Physica D: Nonlinear Phenomena, 215–269, 1999

  11. [19]

    A modified-Leray-α subgrid scale model of turbulence, Nonlinearity, v.19, p.879, 2006

    Ilyin, A.; Lunasin, E.M.; Titi, E.S. A modified-Leray-α subgrid scale model of turbulence, Nonlinearity, v.19, p.879, 2006

  12. [20]

    A data assimilation algorithm for the subcritical surface quasi-geostrophic equation, Advanced Nonlinear Studies, v.14, 167-192, 2017

    Jolly, M.S.; Martinez, V.R.; Titi, E.S. A data assimilation algorithm for the subcritical surface quasi-geostrophic equation, Advanced Nonlinear Studies, v.14, 167-192, 2017

  13. [21]

    A determining form for the damped driven nonlinear Schrödinger equation — Fourier modes case, Journal of Dif- ferential Equations, v.258, 2711-2744, 2015

    Jolly, M.S.; Sadigov, T.; Titi, E.S. A determining form for the damped driven nonlinear Schrödinger equation — Fourier modes case, Journal of Dif- ferential Equations, v.258, 2711-2744, 2015

  14. [22]

    Continuous data assimilation with blurred-in-time measurements of the surface quasi-geostrophic equation, Chinese Annals of Mathematics: Series B, v.40, 721-764, 2019

    Jolly, M.S.; Martinez, V.R.; Olson, E.J.; Titi, E.S. Continuous data assimilation with blurred-in-time measurements of the surface quasi-geostrophic equation, Chinese Annals of Mathematics: Series B, v.40, 721-764, 2019

  15. [23]

    On a well-posed turbulence model, Discrete and Continuous Dynamical Systems series B, v.6, p.111, 2006

    Layton, W.; Lew andowski, R. On a well-posed turbulence model, Discrete and Continuous Dynamical Systems series B, v.6, p.111, 2006

  16. [24]

    Essai sur le mouvement d’un fluide visqueux emplissant l’space, Acta Math, v.63, 193–248, 1934

    Leray, J. Essai sur le mouvement d’un fluide visqueux emplissant l’space, Acta Math, v.63, 193–248, 1934

  17. [25]

    Quelques méthodes de résolution des problemes aux limites non linéaires, Dunod Paris, 1969

    Lions, J.L. Quelques méthodes de résolution des problemes aux limites non linéaires, Dunod Paris, 1969

  18. [26]

    Continuous data assimilation for the three-dimensional Brinkman-Forchheimer-extended Darcy model, Non- linearity, v.29, p.1292, 2016

    Markowich, P.A.; Titi, E.S.; Trabelsi, S. Continuous data assimilation for the three-dimensional Brinkman-Forchheimer-extended Darcy model, Non- linearity, v.29, p.1292, 2016

  19. [27]

    Navier–Stokes Equations and Nonlinear Functional Analysis, So- ciety for Industrial and Applied Mathematics, 1995

    Temam R. Navier–Stokes Equations and Nonlinear Functional Analysis, So- ciety for Industrial and Applied Mathematics, 1995. Departamento Acadêmico de Matemática, Universidade Tecnológica Federal do Paraná, Cornélio Procópio PR, 86300-000, Brazil Email address: deboraalbanez@ut...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.